Flexible Backup Storage Options for Veeam environments with HPE


The Perfect Marriage for Data Protection

One of the best things about Veeam is that, as their tagline suggests, it just works. Beyond the obvious benefits a simpler administrative experience brings, the platform and its foundational components are incredibly flexible in helping meet a wide range of requirements for customers large and small.

Need to reduce load on the production infrastructure and control how data is read from the source? How about minimizing data transmission during the backup window to work around a congested network? What if our backup storage is sub-optimal and we need to make best use of its available performance and capacity? All these elements can be controlled through software with Veeam. It's simple.

Likewise, HPE offers many viable storage options that can be employed successfully within the context of a Veeam architecture. Whether you need a straightforward and reliable x86 server with a large amount of internal storage, a secondary-flash array capable of both housing backup data and running active workloads, or a highly disk-efficient deduplicating storage appliance, HPE has you covered from all angles.

In this post, we'll be taking a look at three example storage options from HPE and analyzing a few of the key use cases for them within a Veeam environment. Although a thorough requirements analysis should take place before a backup storage platform is selected, we hope the examples given here will give you a better idea of which HPE storage solution might work best in your backup environment.

Option 1 - Standard x86 Server – HPE DL380 Gen10 12xLFF 

The HPE DL380 Gen10 is a staple of modern datacenters worldwide, and for good reason. These units are manufactured with high-quality components and are designed with resiliency, performance and scalability in mind.

The 2U formfactor provides 2 CPU sockets and 24 memory slots providing a maximum capacity of 56 cores and 1.5 TB of memory per-socket. This quantity of compute resources is more than adequate for most use-cases, including the ones we are discussing today.

More relevant to the role of backup repository, though, are the IO and storage capabilities of the server. With a total of 12 large form factor drive slots, each server can provide 140TB of usable capacity when populated with 14TB drives.

Using native Veeam software compression and deduplication, it's reasonable to expect up to 280TB usable capacity in this configuration, conservatively assuming a 2:1 data optimization ratio. That's enough capacity for an average customer with an 100TB dataset to retain two weeks' worth of daily backups – all in a simple to administer single-server deployment.

Should more storage capacity be required at a later point, additional disk shelves can easily be added and used to vertically scale the capabilities of the Veeam backup repository server. If it is found that compute and network capacity are also needed in addition to storage, another DL380 Gen10 can be added and a Scale-Out repository can be created.

This has the added benefit of reducing the fault domain should a single backup repository experience a failure. Distribution of risk, especially in a critical area of the design like backup, is a good thing.

Generally-speaking, if you can meet your backup storage requirements using a single (or multiple) DL380's, it's not a bad approach to take. Performance and capacity are more than adequate for many deployments, and the benefits gained from a simpler deployment are significant.

Viable Use Cases
  • Standalone repository
  • Scale-out repository extent
  • Offsite backup copy destination*
Benefits
  • Broad usability in several different deployment roles for a wide range of customer sizes
  • Lowest cost and a minimum amount of administrative overhead and complexity
Identifying Question(s)
  • Is the customer looking for a straightforward backup storage solution that is easy to administer and maintain?
  • Are additional features offered by dedicated secondary storage arrays or deduplication appliances not required?
HPE ProLiant DL380 Gen10

Option 2 - Secondary Flash Array – HPE Nimble SF100

Recent advancements in flash media technology have changed how production workloads, particularly those that are performance-intensive, are stored in the datacenter and cloud. One successful product line made possible by these advancements is the HPE Nimble series of storage arrays.

Designed from the ground up to optimally leverage flash media and equipped with native in-line deduplication capabilities, these systems offer both a high level of performance and data efficiency. Now that the cost-per-gigabyte for flash arrays (both in all-flash and hybrid form) has decreased significantly, these systems are now viable for use as backup storage targets.

Use of a high-performance flash-accelerated backup repository like the HPE Nimble SF100 allows backup data to be used for much more than an occasional test or emergency restore. Production-identical test environments can be easily created from backup data and run directly from the backup storage array, eliminating manual provisioning and configuration overhead and preventing impact on production infrastructure.

These performance characteristics also allow virtual machines to be instantly recovered, mounted to production virtualization hosts and run from backup storage, circumventing lengthy recovery processes. All it takes to finalize the recovery in these scenarios is a simple Storage vMotion back to production storage at an opportune time, like a scheduled maintenance window or other off-peak period.

Fortunately, these additional features do not detract from the core capabilities you'd expect in a backup storage repository. The HPE Nimble SF100 can be equipped with 100TB of usable disk capacity in scaled-up form, which results in a total of 500TB of effective capacity at a 5:1 deduplication ratio.

Using the same 100TB dataset from our fictional customer environment, we can retain up to 90 days' worth of daily backups using less physical disk capacity than the server-based solution. If additional capacity is required beyond what a single scaled-up array can provide, multiple arrays can be combined in a scale-out backup repository configuration, just as with the other options.

Should your production infrastructure also leverage Nimble storage arrays, native storage snapshots can easily be replicated from primary to secondary datacenters and used to supplement your Veeam backups. All of this results in a very capable, efficient, scalable and easy-to-administer solution that leaves little to be desired.

Viable Use Cases
  • Standalone repository
  • Scale-out repository extent
  • Off-site backup copy destination
Benefits
  • Reduction of backup window due to improve ingest / write speed
  • Decrease in actual disk capacity required as a result of efficient array-level deduplication
  • Leverage Instant VM Recovery capability for multiple VM's simultaneously
  • Increase in the speed at which SureBackup (recoverability verification jobs) are performed
Identifying Question(s)
  • Is a large amount of effort currently expended on creating production-identical environments for testing and development purposes?
  • Are secondary copies of production data frequently used for analytics or reporting purposes?
  • Does the presence of very large and/or particularly IO intensive virtual machines complicate the existing recovery strategy?
HPE Nimble SF100

Option 3 - Deduplicating Storage Appliance – HPE StoreOnce 5250  

Technological advancements in the area of backup storage aren't limited to the media data is stored on, either. Purpose-built deduplication appliances like the StoreOnce 5250 leverage advanced data optimization algorithms to achieve very high levels of data deduplication, often up to 20:1 or more with certain datasets.

These results are made possible because deduplication appliances aren't attempting to fill multiple roles – the embedded software is tuned to provide a maximum level of backup capacity using a minimum amount of physical disk. In these devices, performance of sequential reads and writes is emphasized, as this is the workload profile most often seen with backup and restore operations.

General purpose (and secondary) storage arrays that leverage deduplication, on the other hand, must compromise data efficiency for the sake of performance because they are expected to handle a more mixed workload.

With that said, the StoreOnce 5250 is a great option within the StoreOnce line with broad appeal due to its wide range of scalability. It also stacks up well against our previous examples from a usable capacity standpoint. The base unit is equipped with 72TB of usable capacity when populated with 8TB drives, and assuming a conservative deduplication ratio of 10:1, customers can expect a total of 720TB of effective capacity. These figures can be scaled to upwards of 17.3PB effective capacity in a single unit using only five expansion shelves!

If our example customer with a 100TB dataset were to leverage one of these devices as a primary backup target, they could expect to retain more than a year of daily restore points with weekly full backups. As the target for an offsite Veeam backup copy job, a more ideal scenario for this type of solution, our same customer could retain at least seven yearly, eight quarterly, twelve monthly and six weekly backups. That amounts to a very significant amount of data retention in a relatively small footprint, and hopefully a corresponding amount of peace-of-mind for the customer.

If you are looking for a maximum amount of usable backup storage capacity with a minimum of physical disk (and associated datacenter real-estate footprint) required, these solutions are a great choice. Just be mindful to use these devices for their intended purpose – saving the day after a data-impacting event.

Viable Use Cases
  • On or off-site backup copy destination
  • Standalone backup repository*
  • Scale-out repository extent*
Benefits
  • High data ingest speed as a result of optimization for sequential workloads
  • Maximum data efficiency using a minimum of physical disk and datacenter space
  • Efficient data transfer between appliances when using native Catalyst Copy functionality
Identifying Question(s)
  • Does the total amount of data present combined with data retention requirements make use of standard solutions difficult?
HPE StoreOnce 5250

Conclusion

As mentioned previously, a flexible architecture is very desirable when it comes to backup and recovery. With the combination of Veeam backup software and an HPE-based backup repository, customers have just that. With that said, it's no surprise that the two organizations have formed a partnership and have invested in mutual R/D for the benefit of customers.

We hope the examples outlined here are helpful when determining which type of backup repository to leverage within your Veeam architecture. For more detailed architecture and configuration recommendations, be sure consult the Backup Repository section of the Veeam Best Practices site.

And, of course, if you need assistance with assembling your data protection strategy and backup architecture, the experts at IT Partners are always here to help!

HPE Primera and InfoSight
What is a Contract Management Partner?
 

Comments

No comments made yet. Be the first to submit a comment
Guest
Wednesday, 13 November 2019